Abstract
Objective The increased rate of suicide associated with epilepsy has been described, but no studies have reported the rates of suicide and suicide-related behavior (SRB) associated with psychogenic nonepileptic seizures (PNESs).
Methods This retrospective cohort study analyzed data from October 2002 to October 2017 within Veterans Health Administration services. Of 801,734 veterans, 0.09% had PNES, 1.37% had epilepsy, and 98.5% had no documented seizures. Veterans coded for completed suicide, suicide attempts, and suicidal ideation were identified from electronic health records. The primary measure was the suicide-specific standardized mortality ratio (SMR) based on the number of suicide deaths and CDC national suicide mortality database. A Poisson regression was used to calculate the relative risk (RR) of suicide across groups.
Results A total of 1,870 veterans (mean age [SD] 33.76 [7.81] years) completed suicide. Veterans with PNES (RR = 1.75, 95% confidence interval [CI] 0.84–4.24) and veterans with epilepsy (RR = 2.19, 95% CI 2.10–2.28) had a higher risk of suicide compared with the general veteran population. Veterans with PNES or epilepsy had a higher risk of suicide and SRB if they had comorbid alcohol abuse, illicit drug abuse, major depression, posttraumatic stress disorder, and use of psychotropic medications. Conversely, those who were married or attained higher education were at a decreased risk. The SMR for completed suicide for PNES, epilepsy, and the comparison group was 2.65 (95% CI 1.95–5.52), 2.04 (95% CI 1.60–2.55), and 0.70 (95% CI 0.67–0.74), respectively.
Conclusions Veterans with seizures (both psychogenic and epileptic) are at a greater risk of death by suicide and SRB than the comparison group. These findings suggest that although the pathophysiology of PNES and epilepsy is different, the negative impact of seizures is evident in the psychosocial outcomes in both groups.
Suicide is a common cause of increased mortality in people with seizures, and the relationship between seizures and suicide is complex and controversial.1,2 Antiseizure medications (ASMs) have a black box warning for increased risk for suicide.3 However, several large cohort studies demonstrate that epilepsy increases the risk of completed suicide and suicide-related behavior (SRB) independent of ASMs, even after successful epilepsy surgery.4,-,6 Changes in neural networks may make people with epilepsy more vulnerable to mood disorders.7,8 However, the relationship of psychosocial stressors related to seizures and suicide risk is understudied. For instance, social stigma and isolation, driving restrictions, and decreased employability may play a major role in suicide risk. People with psychogenic nonepileptic seizures (PNESs) face similar psychosocial challenges to people with epilepsy, and most carry the diagnosis of epilepsy for years before the PNES diagnosis. However, they do not have the same neuropathology as people with epilepsy.9 Therefore, comparing the suicide rate of people with epilepsy to people with PNES may provide insight into the role of psychosocial forces in suicide and SRB.
Methods
SRB was defined as the presence of suicidal ideation and/or suicide attempts as per International Classification of Diseases (ICD)-9 and -10 coding (see eTable 1, links.lww.com/CPJ/A271, for ICD-9 and -10 codes used). Groups were divided into veterans coded for PNES, veterans coded for epilepsy, and the general veteran group composed of veterans without coding for PNES or epilepsy.2,10 People with combined PNES and epilepsy were removed from the analysis. The sample (n = 801,734) was drawn from an existing cohort study that includes men and women veterans enrolled in Veterans Health Administration (VHA) care from October 2002 to 2017.10 A validated set of ICD-9 and -10 codes for intentional or undetermined violent deaths (defined as completed suicide) was obtained from the VA Suicide Data Repository.11,12
Standard Protocol Approvals, Registrations, and Patient Consents
The protocol was approved by VA Connecticut Healthcare System Institutional Review Board as well as The U.S. Army Medical Research and Development Command Office of Research Protection Office.
Statistical Analysis
Standardized mortality ratio (SMR) was calculated to compare the suicide mortality rates among veterans with PNES, veterans with epilepsy, and the general veteran group in comparison to expected suicide deaths over a 16-year period. Data were stratified by sex (female/male) and age (15–24, 25–34, 35–44, 45–54, 55–64, 65–74, and 75–84 years old). Data for the standardized mortality ratio were obtained from the CDC (Matthew M. Zack, CDC medical epidemiologist, email communication, February 11, 2020).
Chi‐square tests were used to examine bivariate relationships between categorical variables, and t tests or nonparametric tests, as appropriate, were used for continuous variables. Because our study is a cohort study and the primary outcomes were rare events over time, we used Poisson regression with a log link and robust variance estimates to calculate relative rates of suicide and SRB and 95% confidence intervals (CIs). The index date used for SRB for veterans with PNES and epilepsy was the first incident recorded, as for suicide it is a 1-time event. Demographic factors include sex and race. The covariates are alcohol abuse, drug abuse, major depression, posttraumatic stress disorder (PTSD), psychotropic medications, educated, and married. Traumatic brain injury (TBI) and anxiety as covariates were removed from the final Poisson model as there was no statistical significance. To control for potential confounding that could distort the risk estimates, demographic clinical variables that were significant in bivariate analyses were entered into the model as covariates. An alpha level of 0.05 was considered statistically significant. All statistics were conducted using SAS version 9.4 (Cary, NC).
Data Availability
The data generated and/or analyzed during the current study are not publicly available due to VA privacy and information security policies. Deidentified data may be requested and shared with the appropriate permissions and approvals.
Results
The cohort included a total of 801,734 veterans: 752 veterans with PNES, 10,994 veterans with epilepsy, and 789,988 veterans without PNES or epilepsy (defined as other) (Table 1). The mean age was 31.85 years (SD 9.07) at the first visit and 33.76 years (SD 7.81) for suicide. There were 1,870 total deaths from suicide: 0.80% veterans with PNES, 0.65% of veterans with epilepsy, and 0.23% of veterans in the general veteran group.
Demographics for All Deaths in Veterans (n = 801,734, Mean Age = 31.8 Years) With PNES, Epilepsy, and Veterans in the General Veteran Group (Identified as Other)
SRB was present in 29% of veterans with PNES, 24% of veterans with epilepsy, and 5.6% of the general veteran group. The SMR was calculated using CDC national suicide mortality data to standardize for age and sex. The SMR for PNES, epilepsy, and the general veteran group was 2.65 (95% CI 1.95–5.52), 2.04 (95% CI 1.60–2.55), and 0.7 (95% CI 0.673–0.738), respectively (Figure 1). The timing of suicide ideation and attempts were well distributed around the date of epilepsy and PNES diagnosis; however, >80% of deaths due to suicide among people with epilepsy occurred 1 year after diagnosis (eFigure 1, links.lww.com/CPJ/A270 and eAppendix 1, links.lww.com/CPJ/A269).
The SMR was 2.65 (95% CI 1.95–5.52) for veterans with PNES, 2.04 (95% CI 1.60–2.55) for veterans with epilepsy, and 0.7 for veterans in the general veteran group (95% CI 0.673–0.738). CI = confidence interval; PNES = psychogenic nonepileptic seizure.
The relative rates for suicide and SRB from Poisson regression are shown in Table 2. Veterans diagnosed with PNES have an RR of 1.75 (95% CI 1.51–2.02; p < 0.0001) times that of epileptic veterans for committing suicide and an RR of 3.82 (95% CI 3.32–4.39; p < 0.0001) times more likely to commit suicide compared with the general population. Epileptic veterans have an RR of 2.19 (95% CI 2.10–2.28; p < 0.0001) times the risk of suicide as opposed to the other group or general population. For those veterans with SRB, it is reported that veterans with PNES have an RR of 1.61 (95% CI 1.51–1.73; p < 0.0001) compared with epileptic veterans and an RR of 3.65 (95% CI 3.42–3.90; p < 0.0001) compared with the general population (other group). Veterans with epilepsy are 2.26 (95% CI 2.22–2.31; p < 0.0001) times more likely to report SRB than the general population. Female veterans with PNES are 0.48 (95% CI 0.46–0.50; p < 0.0001) times as high a risk of committing suicide compared with their male counterparts and 0.96 (95% CI 0.95–0.98; p < 0.0001) times as high a risk for engaging in SRB. Of the racial group, veterans who classified as other compared with White veterans were at a higher RR 0.81 (95% CI 0.79–0.83; p < 0.0001) of committing suicide compared with Black 0.45 (95% CI 0.44–0.46; p < 0.0001) and Hispanic 0.56 (95% CI 0.55–0.58; p < 0.0001) veterans. For those veterans with PNES engaging in SRB, the Black population has the greatest risk with an RR of 0.93 (95% CI 0.92–0.94; p < 0.0001). Anxiety and TBI were significantly associated with SRB only (p < 0.0001). Of comorbid covariates, veterans with PNES have the greatest risk of suicide if they are engaging in drug abuse with an RR of 1.92 (95% CI 1.88–1.96; p < 0.0001). Veterans with PNES and who are on psychotropic medication have the highest risk of engaging in SRB with an RR of 8.38 (95% CI 8.15–8.61; p < 0.0001).
Poisson Regression of Veterans With Seizures (PNES and Epilepsy) and Associated Risk Factors Compared With the General Veteran Group
Discussion
Veterans with epileptic and nonepileptic seizures are both approximately 2 times more likely to commit suicide and exhibit SRB than veterans in the general veteran group (Figure 1). Increased risk of SRB has been attributed to ASMs, psychiatric comorbidities, psychosocial stressors, and changes in neurocircuitry. The underlying neurobiology of epilepsy and PNES fundamentally differs. Therefore, the high risk of suicide and SRB in both populations is likely not due to a common neurobiological substrate but may be due to the shared psychosocial stressors of unpredictable seizures and high levels of stigma.
Mood disorders and substance abuse are strongly associated with suicidality. The diagnosis of a mood disorder is associated with a 35%–60% increased risk of suicide and over 60% of those who commit suicide were previously diagnosed with a mood disorder.13 Our study demonstrated similar increased risk with depression, psychotropic medication use, alcohol abuse, and drug abuse. As reviewed elsewhere, education and marriage are associated with reduced risk of suicide.14,15 TBI and anxiety were not significantly associated with completed suicide, but the association became significant for SRB. However, it is important to note that these risk factors are not mutually exclusive, as TBI could cause depression and anxiety, and PTSD could drive alcohol and drug abuse.
PTSD is an important risk factor for suicidality and mental health comorbidities. Although PTSD is a confounder to suicide and SRB, it does not fully explain the increased suicide seen in PNES and epilepsy. When analyses were controlled for PTSD, the association between suicide and seizures was still significant. Thus, higher rates of PTSD only partially explain the increased rates of suicide among veterans with epilepsy and PNES.
The decreased SMR in the general veteran group (SMR = 0.7) is consistent with current literature. Hoffmire and Bossarte16 reported that veterans who died of suicide were older than nonveterans who died of suicide. Veterans who use VHA services may have a decreased suicide risk compared with those who do not.17
Further research should explore the association of psychosocial risks and protective factors in people with seizures and suicidality. PNES is a functional neurologic disorder that is diagnosed by neurologists and treated by mental health professionals, yet many patients still have poor outcomes as demonstrated by the increased suicide risk.
External validity to the general US population may be limited given that our study was restricted to patients receiving primary care through the VA, who are more likely to be men and have more comorbidities than counterparts receiving care outside the VA. The majority of our population was between 30 and 49 years old and male due to the inclusion criteria of the database we used.
We used administrative databases (ICD-9 and ICD-10); therefore, each patient was not systematically evaluated. Because of the limited sample size for suicide in PNES, the CIs are wide for suicide and SRB.
Veterans with PNES and veterans with epilepsy are both at significantly increased risk for suicide and SRB when compared with the general population. The high risk of suicide among veterans with different types of seizures may be due to shared psychosocial stressors. Future research will look at whether PTSD therapy and PNES-driven psychotherapy is protective for suicide and whether delayed diagnosis of PNES plays a part in the increased suicidality.
Acknowledgment
The authors thank members of the VA Connecticut Healthcare System, the Pain, Research, Informatics, Medical comorbidities and Education Center of Innovation, and the Yale Center for Neuroepidemiology and Clinical Neurological Research for their input and feedback.
Study Funding
Study funded by the Department of Defense Congressionally Directed Medical Research Programs Epilepsy Research Program (DOD CDMRP ERP) #EP160049.
Disclosure
Y. Bornovski's salary was paid by a grant from the US Department of Defense CDRMP (EP1600449). E. Jackson-Shaheed, S. Argraves, and A. Hitchins report no disclosures relevant to the manuscript. B. Tolchin received research funding from a US Veteran Administration (VA) VISN1 Career Development Award, the VA Pain Research, Informatics, Multimorbidities, and Education (PRIME) Center of Innovation, and the C.G. Swebilius Trust; he has received honoraria from the Columbia University Medical Center, the International League against Epilepsy, and the American Academy of Neurology. D. Galluzzo, K.-H. Cheung, J. Goulet, M. Skanderson, C.A. Brandt, and M.J. Pugh report no disclosures relevant to the manuscript. H. Altalib is on the scientific advisory board for Eisai; has received commercial research support from Eisai Medical Research, Engage Therapeutics, Pfizer, Sunovion, and UCB; and has received government research support from the Department of Defense Congressionally Directed Medical Research Programs Epilepsy Research Program (DOD CDMRP ERP) #EP160049, DOD CDMRP ERP #EP150013, Department of Veterans Affairs Health Service Research & Development Center of Innovation # CIN 13-407 COIN, and US Department of Defense Office of Naval Research #N66001-08-C-2005. Full disclosure form information provided by the authors is available with the full text of this article at Neurology.org/cp.
Appendix Authors
Footnotes
Funding information and disclosures are provided at the end of the article. Full disclosure form information provided by the authors is available with the full text of this article at Neurology.org/cp.
- Received July 10, 2020.
- Accepted January 18, 2021.
- © 2021 American Academy of Neurology
The Nerve!: Rapid online correspondence
REQUIREMENTS
If you are uploading a letter concerning an article:
You must have updated your disclosures within six months: http://submit.neurology.org
Your co-authors must send a completed Publishing Agreement Form to Neurology Staff (not necessary for the lead/corresponding author as the form below will suffice) before you upload your comment.
If you are responding to a comment that was written about an article you originally authored:
You (and co-authors) do not need to fill out forms or check disclosures as author forms are still valid
and apply to letter.
Submission specifications:
- Submissions must be < 200 words with < 5 references. Reference 1 must be the article on which you are commenting.
- Submissions should not have more than 5 authors. (Exception: original author replies can include all original authors of the article)
- Submit only on articles published within 6 months of issue date.
- Do not be redundant. Read any comments already posted on the article prior to submission.
- Submitted comments are subject to editing and editor review prior to posting.